Research Article

Flux and composition of interstellar dust at Saturn from Cassini’s Cosmic Dust Analyzer

See allHide authors and affiliations

Science  15 Apr 2016:
Vol. 352, Issue 6283, pp. 312-318
DOI: 10.1126/science.aac6397

Cassini detects interstellar dust grains

The interstellar medium contains an array of small solid particles known as dust grains. Altobelli et al. used the dust analyzer on the Cassini probe to detect 36 interstellar dust grains as they passed by Saturn, and they measured the grains' elemental abundances. The results show that, remarkably, these grains lack carbon-bearing compounds and have been homogenized in the interstellar medium into silicates with iron inclusions.

Science, this issue p. 312

Abstract

Interstellar dust (ISD) is the condensed phase of the interstellar medium. In situ data from the Cosmic Dust Analyzer on board the Cassini spacecraft reveal that the Saturnian system is passed by ISD grains from our immediate interstellar neighborhood, the local interstellar cloud. We determine the mass distribution of 36 interstellar grains, their elemental composition, and a lower limit for the ISD flux at Saturn. Mass spectra and grain dynamics suggest the presence of magnesium-rich grains of silicate and oxide composition, partly with iron inclusions. Major rock-forming elements (magnesium, silicon, iron, and calcium) are present in cosmic abundances, with only small grain-to-grain variations, but sulfur and carbon are depleted. The ISD grains in the solar neighborhood appear to be homogenized, likely by repeated processing in the interstellar medium.

Most of the presolar dust grains, containing the heavy elements from which the planets formed, were destroyed or heavily processed during the first stage of solar system formation. Although a minor population of presolar grains, recognized by their extremely diverse isotopic composition, survived this process in meteorite parent bodies (1), it is completely unknown whether those grains are representative of the grain populations in the interstellar medium (ISM) (24). Interstellar dust (ISD) from the local interstellar cloud (LIC) is injected into the solar system as a highly directional stream nearly aligned with the ecliptic (heliocentric ecliptic longitude/latitude of 79°/–8°) (2, 5). This LIC-ISD flow was first detected in situ by the dust instrument onboard the Ulysses spacecraft (6) and has been monitored using a variety of dust detectors at solar distances ranging from 0.3 to 5 astronomical units (AU) (5, 711). In situ sampling of the ISD flow was attempted by the Stardust mission, and recent laboratory analyses of the aerogel collectors have revealed three tracks caused by ISD impactors, of which two showed an ISD end particle and the third residuals on the track walls. Elemental residues were also found, probably of ISD origin, in four impact craters in the aluminum foil collectors (12). The Cassini spacecraft has been in orbit around Saturn since 2004, carrying the Cosmic Dust Analyzer (CDA), whose chemical analyzer (CA) subsystem acquires time-of-flight (TOF) mass spectra of dust particles, revealing the grains’ elemental composition (13). Primarily designed to study the dust populations native to the Saturnian system, Cassini CDA has also proven to be an in situ observatory for dust from beyond Saturn. In this work, we present the result of the analysis of 10 years of Cassini CDA data (mid-2004 until 2013), revealing the faint but distinct signature of LIC-ISD in the largely dominant background of Saturn’s E ring water-ice particles originating from Enceladus (14).

Flux, mass distribution, direction, and dynamics of ISD grains at Saturn

We identify a total of 36 ISD grains within a distance of 9 to 60 Saturn radii from the planet, characterized by their high entry speed into the Saturnian system, from a direction compatible with the expected LIC-ISD flow at Saturn (Fig. 1). Both the cumulative radial and temporal distribution of the ISD grain detections at Saturn can be reproduced if the ISD grains in the observed mass range account for an average flux of 1.5 × 10−4 m−2 s−1, corrected for Saturn’s motion (Fig. 2) (see the supplementary materials). The time period covered by our analysis was particularly favorable to ISD detection at Saturn. From 2004 onward, Saturn was moving toward the ISD downstream direction, gradually increasing the relative flux, which peaked in 2010 when Saturn’s velocity vector direction was aligned to the ISD upstream direction (15). Moreover, in 2010 there was a strong focusing of the smaller particles by the Lorentz force resulting from the relative particle motion with the interplanetary magnetic field (IMF) (16). The ISD stream is well collimated around the reference ISD direction, within the uncertainty of the impact direction measurement of 28° (Fig. 2 and supplementary materials). From the measured impact ion charge amplitude and modeling of the grain dynamics, we derived the mass as well as the radiation pressure-to-gravity ratio (β) for each grain (see the supplementary materials). The method used allows for a mass derivation independent of any assumption on β. The mass distribution and β ratio are plotted on Fig. 3. The mass distribution peak is located at ~10−17 kg, with a decrease in grain abundance above 10−16 kg and a sharp cutoff below 5 × 10−18 kg. The deficit of large grains (above 10−16 kg) is due to the lack of sensitivity of our ISD identification methods based on the TOF spectra, preventing us from detecting the largest ISD grains present in the Ulysses data (17, 18).

Fig. 1 Minimum speeds inferred from TOF mass spectra.

Plotted are the minimum speeds of nonwater dust grains coming from within 50° of the reference ISD direction (see the methods), with respect to Saturn as a function of the radial distance to Saturn. The color code indicates the total charge of the ions (ion yield) generated upon impact. The grains identified as ISD are plotted using an asterisk symbol. The plain and dashed lines in black indicate reference speeds with respect to Saturn as a function of the radial distance: the speed on Keplerian bound circular orbits, the spacecraft speed averaged per distance bins over 2004 to 2013 (SC vel), and the theoretical speed of interplanetary particles, depending on their injection speed at Saturn’s Hill sphere (vinj). (For comparison, see fig. S3.)

Fig. 2 ISD flux and direction.

Cumulative temporal and radial distance distribution of the ISD grain detections (black dotted-dashed lines on the upper and middle panels, respectively). The plain lines are the modeled distributions obtained by computing the number of ISD grains from the reference ISD direction with a flux of 1.5 × 10−4 m−2 s−1 (see the methods section of the supplementary materials). The lower panel shows the distribution of the angles between the CDA boresight at the time of ISD grain detection and the ISD reference direction. The relative time spent in each angle bin over the time period 2004 to 2013 is indicated by the dashed line. Although the CDA spends most of the time pointing 30 or 40° away from the reference ISD direction, all except one ISD grain were detected within 20° of the reference ISD direction.

Fig. 3 ISD grain mass distribution and β-mass values.

The dashed black line indicates the grain mass distribution. The individual β and mass values inferred for each ISD grain are plotted with error bars. A one order of magnitude error in the mass determination results from the mass-velocity calibration uncertainty (13). A maximum error of 0.2 in the determination of β is applied for all grains. This maximum error takes into account a range of possible (β-dependent) velocity values. Theoretical β-mass values obtained using Mie theory computations for amorphous grains with different compositions and structures are indicated as plain and dashed lines. MgSiO3 only refers to composition and does not imply crystallinity. The gray rectangular areas show the range of β-mass values inferred from the Ulysses ISD data (3). A structure with heterogeneous distribution of Fe and/or C is needed to explain the relatively high β values.

Composition of ISD grains

Figure 4 (upper panel) shows a typical mass spectrum from an ISD grain. Owing to the high velocity of ISD impacts (typically above 20 km/s), the energy densities upon impact are high enough to totally vaporize the solid grain and yield cation spectra that are clearly dominated by elemental ions rather than molecular ones. The main elemental cation peaks in all ISD spectra (going from lower to higher atomic masses) are carbon, oxygen, sodium/magnesium, potassium/calcium, iron, and rhodium. Rhodium, however, is not indigenous to particles but is excavated from CA target material, as well as a substantial fraction of the observed carbon that is contamination from the Rh target (19) (see the supplementary materials). Owing to the relatively low mass resolution of the CDA, signatures of Na/Mg and K/Ca cannot be easily separated in most spectra. However, the positions of the maxima of the respective peaks indicate that Mg+ and Ca+ dominate the adjacent species in almost every ISD spectrum. Although not forming individual peaks, Al+ and Si+ are visible as an extended flank on the Mg signature. In one individual spectrum, cations of Cr are indicated. The Fe signature is very broad in every ISD spectrum; therefore, although not detectable, minor amounts of Ti, V, Mn, and Ni are possible. We use other nonwater spectra of particles from the ISD direction, which, however, do not fulfill the speed criterion for ISD selection (see Fig. 1), for comparison. These particles are hereafter called the reference group (see the supplementary materials). We compare on Fig. 4 (lower panels) the Mg+/Fe+ line amplitude ratios and Mg+/Ca+ line amplitude ratios of ISD grains with the reference group. The compositional difference between the ISD and the reference group is apparent at first glance. The ISD grains (i) are clearly compositionally distinct from the reference group and (ii) exhibit a very similar composition within their own group.

Fig. 4 Typical CDA ISD TOF mass spectrum and comparison of ISD spectra with reference group spectra.

The upper panel shows a typical impact ionization mass spectrum from an ISD grain. Vertical lines indicate the positions of relevant cation species. The x axis indicates the TOF needed by the cations to reach the ion detector (after triggering the event), from which the ion masses are derived. The y axis shows the amplitude of the electrical signal derived from the detected cations, a measure of their abundance. The ions of the main particle constituents form the main peaks: Mg, Fe, Ca, and O. Si and Al are forming a distinct shoulder on the right flank of the Mg signature. The Rh peak is from the CDA’s impact target made of rhodium. To convert the ion abundance into elemental abundance, RSFs have to be applied (see the methods). In the lower panel are shown the ratios of cations observed in ISD spectra (upper panels) and spectra of the reference group (lower panels). The narrow distributions observed in ISD spectra suggest that ISD grains have a similar composition.

To quantify the relative contributions of the major elements Mg, Si, Ca, and Fe in individual ISD particles, we applied the following procedure (for details, see the methods section of the supplementary materials): For each individual mass spectrum, a model spectrum was calculated using a least-square fit procedure to obtain the relative contributions of ions of different elements, particularly to deconvolve the Si peak from the Mg peak tail. Because all elements have a different probability to form cations (e.g., Mg forms cations five times more often than silicon), ion abundances were translated into element abundances by using appropriate relative sensitivity factors (RSFs) (supplementary materials) (20). Resulting element composition ratios are shown in Fig. 5 [ratios by weight % (wt %)]. It can be verified that Mg/Si, Fe/Si, Mg/Fe, and Ca/Fe ratios are on average CI chondritic (carbonaceous chondrite of type Ivuna whose composition is considered as a proxy of solar or cosmic element abundances) and similar to a composition inferred for LIC-ISM dust (21). For comparison, Fig. 5 (upper panels) shows both the range of elemental compositions for our ISD data and isotopically anomalous presolar stardust grains found in meteorites (2228). Those grains condensed in the outflows of asymptotic giant branch (AGB) stars and supernovae and became part of the ISM before they were incorporated into the parental molecular cloud from which our solar system formed.

Fig. 5 Comparison of element wt % ratios in CDA ISD grains with chondritic and presolar material.

Mg/Si, Fe/Si (lower panel, left side) and Mg/Fe, Ca/Fe ratios (lower panel, right side) are shown for ISD grains (blue squares). The elemental ratios given by the x axis of each lower panel also apply to the upper panels. The average composition (large blue square) is compatible with CI chondrites (yellow triangle), indicating average cosmic element abundance ratios, and is also compatible with LIC-ISD composition from astronomical observations (green triangle). Limited compositional variation is possible, but at least part of this scatter is due to the detection method. For comparison, we analyzed compositional data of 388 presolar circumstellar dust grains, which were obtained by Auger spectroscopy (2228). Grains of silicate composition [i.e., with atomic ratios of (Fe+Mg+Ca)/Si < 2.2] are shown. Histograms for Mg/Si and Mg/Fe (upper panels) demonstrate their larger scatter and more heterogeneous composition compared with the ISD grains of this study. For low Ca presolar silicate grains (gray diamonds), the Ca/Fe values are upper limits, assuming the detection limit as Ca abundance; hence, the true scatter of the presolar silicate grains in Ca/Fe is actually larger. The smaller scatter of the LIC-ISD grain population in this study indicates that they represent a secondary population that was derived by processing and homogenization in the ISM from a primary, more heterogeneous circumstellar grain population. Also shown are typical compositions of silicates like forsterite (Fo90Fa10), fayalite (Fa90Fo10), and Mg-rich clino- and orthopyroxene (CPx and OPx) from a primitive mantle lherzolite (20), again emphasizing that the LIC-ISD grains represent average compositions rather than mineral end-member compositions. Two data points are shown for the Orion and Hylabrook particles returned by the Stardust mission: Although Mg/Fe is largely compatible with our ISD data, Mg/Si is significantly higher.

No mass line in agreement with sulfur can be observed in any of the ISD spectra. Sulfur is considerably harder ionized than most metals, but it forms abundant cations at impact speeds exceeding ~9 km/s (29). Ascribing all ions detected between mass 31 and 34 to sulfur (although these ions do not form a distinct peak there; see Fig. 4) and using the minimum RSF of S for impact speeds between 20 km/s and 30 km/s normalized to Fe by (29), we calculate an upper limit for the average S/Fe (atomic) ratio of 0.0952 + 0.101/– 0.077. This upper limit is consistent with the value computed for LIC-ISM dust of 0.097 ± 0.058 (30). However, it is significantly lower than the CI chondritic ratio of 0.48.

Carbon is another element that appears to be highly depleted in our ISD mass spectra. A significant carbon peak is present in all spectra, but figs. S1 and S2 show that a greater part of it is due to a well-known target contamination of the CDA (19). Only if we neglect target contamination and ascribe the entire carbon signal to impacting particles and if we furthermore assume very low detection efficiencies of carbon (see the supplementary materials’ section on carbon abundance), a quantification results in an average C/O atomic ratio of 0.32 ± 0.25, which would be consistent with an LIC-ISM dust value of 0.32 ± 0.29 (3). This value assumes a higher fraction of carbon in the gas phase, in contrast to an alternative value of 0.81 ± 0.55, which is calculated based on different solar photospheric abundances (3). It could well be, however, that only 10% or less of the carbon is indigenous to the particles (figs. S1, S2, and S6) and that we underestimate the carbon detection efficiency by an order of magnitude, which would lower the observed C/O ratio to <0.032—i.e., far below the ISM dust value.

Discussion

Our analysis shows that ISD grains from the local interstellar neighborhood of the solar system can reach the inner Saturnian system. Given the relative insensitivity of the CDA’s CA for large grains, it is clear that the mass distribution derived is representative of the smaller grains only up to about 5 × 10−16 kg. For the same reason, our flux value is a lower estimate, and the total flux accounting for the undetected large grains could be up to twice as large (3 × 10−4 m−2 s−1). Our measurements suggest, therefore, that the ISD flux at Saturn (10 AU) is larger than the fluxes measured by Ulysses at distances up to 5 AU. It is likely that this larger ISD flux at Saturn is related to a focusing effect predicted for the grains as they interact with the IMF (16) during our collection period, confirming that the ISD grains, in the size range covered by our study, have a nonzero charge-to-mass (q/m) ratio.

The deficit of very small grains in our data, despite the sensitivity of our method that enables the detection of grains with masses as low as 10−19 kg, can be explained by the filtering of such small grains at the heliopause and inner heliosphere (15, 31, 32). This phenomenon was also observed in the Ulysses ISD data (17). We confirm the existence of grains with β > 1 with a maximum value reached between 10−17 and 10−16 kg, in good agreement with the β-mass domains inferred from Ulysses data (3).

The β-mass curve by itself cannot directly provide the grain composition and structure unambiguously but provides an independent consistency check of the compositional measurements by the CDA mass spectra. The β ratio quantifies the solar photon scattering efficiency and, combined with the grain mass, is a proxy for the grain composition and structure (3, 16). Synthetic β-mass curves were computed using Mie theory for different grain models with bulk composition compatible with the measured TOF spectra and compared with the β-mass curve inferred from the grain dynamics (Fig. 3). All models of grains considered for the β computation are amorphous but compact particles, by opposition to porous structures. For the size regime of less than a few hundred nanometers observed by the CDA, the assumption of compact particles is the most plausible one and is reasonably supported by our observations (see supplementary materials’ section on derivation of the β-mass curve). In contrast, larger ISD in the μm regime are likely porous aggregates (33). Good agreement is found for silicate particles enriched with metallic Fe. In contrast, the β ratio of pure Fe-Mg silicates without metallic Fe enrichment cannot exceed unity, which is in disagreement with the high β ratio of most grains above 5 × 10−17 kg. As we find no indications for abundant organic material in CDA mass spectra, we conclude that at least some of the iron identified by the CDA has to be in metallic form. However, our upper limit for carbon does not exclude the presence of thin organic mantles on the ISD grains causing an increase in the β ratio. Collecting ISD grains at Saturn allowed us to sample LIC dust at the largest heliocentric distance yet achieved in situ, reducing the amount of dynamical filtering as compared with former detections performed closer to the Sun. In addition, the relative position of Saturn and the Sun, with respect to the ISD stream over the time period considered in this work, permitted the detection of grains with a wide range of β values (16). We are therefore convinced that our compositional results are not significantly affected by a selection bias related to the grain dynamics in the solar system.

All models of LIC-ISD grain composition thus far have been based on the ultraviolet (UV) spectra measurements of absorption lines of C, N, O, Mg, Al, Si, S, and Fe in the gas phase of the LIC along the lines of sight toward nearby stars (2, 3, 34). These models assume that the element total abundance corresponds to cosmic abundances derived from the solar photosphere composition and that an element depletion in the gas results from its incorporation into dust grains. Most of these studies suggest dust populations with different compositions: a Mg-rich silicate population, a carbon-rich population, Fe-rich dust (possibly oxides), and an Al-rich group (possibly corundum). A model of ISD grains in the general ISM (35, 36) proposes a core of silicate composition and an organic mantle remaining after UV radiation of ices. A more detailed mineralogical model was proposed (3), with grains made of a core of Mg-rich olivine and pyroxene composition and iron-bearing metallic inclusions (like kamacite), coated by an organic refractory mantle containing CHON-bearing material.

Our in situ detection of ISD provides constraints on these models. First of all, the average composition of our Mg, Si, Ca, and Fe data is in general consistent with the observation that these elements are fully condensed from the gas that forms the LIC, resulting in cosmic (CI chondritic) element abundances (Fig. 5), and that particles of silicate composition dominate the LIC-ISD population. More important, our data demonstrate that compositional homogeneity extends down to small spatial scales of 200 nm—i.e., the grain size of the ISD particles observed by the CDA. All grains contain these silicate-forming elements, and individual grain-to-grain variations are limited (see Figs. 5 and 6; note that some, if not most, of the scatter of our data is due to stochastic uncertainties of the impact ionization of the CDA). In the investigated size regime, there are no indications of other, compositionally diverse ISD populations such as pure Fe-O grains or pure carbonaceous particles. The absence of the latter was already suggested by (34, 37). Although we cannot entirely rule out their existence in the LIC, such populations do not appear to represent more than an upper limit of Embedded Image of the ISD grains observed by the CDA penetrating into the solar system (assuming Poisson statistics for the error on the number of ISD grains detected). The depletion of carbon and sulfur inferred from the gas-absorption spectra of the LIC indicates loss of these compounds from ISD grains. The most plausible explanation is that volatile sulfur and carbon are incompletely condensed under conditions applicable for the LIC. In this respect, it should be noted that Mg/Si and Fe/Si ratios are slightly higher than cosmic abundances and could indicate a small deficit of Si, which is more volatile than the other major rock-forming elements Mg or Fe. The process responsible for the loss of these volatile substances needs further clarification.

Fig. 6 Comparison of (Mg+Fe+Ca)/Si ratio histograms (in atom %) of CDA ISD grains, presolar circumstellar material, LIC-ISD determined by astronomical observations, and ISD returned by the Stardust mission.

Presolar grains have a maximum at a silicate composition intermediate between pyroxene and olivine and a tail to high values caused by a bias (overabundance) of refractory oxides. LIC-ISD values center between 2 and 3. Such high values are caused either by oxides or by metallic Fe, phases that have also been observed in the Stardust particles Orion and Hylabrook. The presolar circumstellar grain population is biased; for example, there is possibly selective loss of amorphous grains or a higher fraction of resistant oxides due to metamorphic destruction on asteroids or comets. Nevertheless, it is highly unlikely that the presolar grain population initially had a distribution similar to our ISD data: In this case, the presolar population should have lost primarily the oxide or metal component, which is, however, unlikely, as oxides are rather more resistant against metamorphic destruction.

Only the Stardust mission has previously provided insights into the composition of contemporary interstellar dust crossing our solar system, with material collected in aerogel and extracted from four impact craters resulting from ISD impacts in the aluminum foil collectors (12, 38, 39). The sizes of these impact craters between 0.28 and 0.46 μm imply particles of similar sizes as those observed with the CDA (12). The inferred bulk particle composition is in agreement with our results, except for the presence of sulfide in three of the four craters (39), whereas there is no indication of sulfur in CDA ISD mass spectra. Three particles were collected in aerogel; two of them—Orion and Hylabrook—yielded compositional data (39) that allow a comparison in Fig. 5. Although Mg/Fe ratios are within the range of the CDA’s ISD particles, the Mg/Si and Fe/Si ratios are significantly higher—i.e., the Stardust ISD particles are depleted in Si with respect to cosmic abundances. It should be kept in mind, however, that Orion and Hylabrook are much larger and perhaps porous particles, so that the formation mechanisms and the particles’ chemical history might not be directly comparable to our data.

Our results can also be compared with elemental compositions of interstellar grains preserved in 4.6-billion-year-old meteorites (Fig. 5). These interstellar grains survived processing in the protoplanetary disc when our solar system formed and consist of compositionally diverse populations. Their specific isotopic compositions reflect the nucleosynthetic processes of their host stars at the time, when these grains condensed in the outflows of the parental AGB stars or supernovae. Analysis of the inventory of presolar grains in meteorites indicates that the main population is silicates (e.g., olivine and pyroxene), with minor contributions (i.e., a few percent) of oxides (e.g., corundum and hibonite) and carbonaceous grains, mainly silicon carbide (40), with an abundance of >20%, possibly up to 50%.

The upper panels of Fig. 5 show histograms of Mg/Si and Mg/Fe ratios of presolar silicate populations and our ISD data. It can be verified that the scatter of elemental compositions of presolar silicates is about a factor of 2 larger (on logarithmic scales) compared with the CDA ISD data. Figure 6 additionally displays (Mg+Fe+Ca)/Si ratios (atom %): Although the presolar grain population has a maximum between pyroxene- and olivine-type silicate composition with values between 1 and 2, our ISD grain population centers between 2 and 3, similar to LIC-ISM dust composition inferred from astronomical observations (37). Such values can be explained by the presence of either Mg, Fe, or Ca oxides or metallic Fe. These phases occur also in the Orion and Hylabrook ISD particles, which also have high (Mg+Fe+Ca)/Si ratios. For the Stardust ISD residues on Al foils, only one particle yielded a quantifiable bulk value of 3.3, whereas for the remaining residues, the bulk values could not be constrained because of the uncertain proportions of silicates and Fe metal and sulfide.

These observations indicate that the major population of dust grains in the ISM is different from pristine circumstellar stardust. A more homogeneous composition concerning Mg/Si or Mg/Fe can be explained by destruction, recondensation, and equilibration processes in the ISM that homogenize an initially diverse population that started as circumstellar dust. This could also explain the observation that the high Mg end of the presolar silicate population in Fig. 5 (which are crystalline circumstellar olivine grains) is missing in the Cassini ISD data: Grain destruction and recondensation will wipe out the high Mg tail and result in a very low fraction of crystalline ISM grains. The same conclusion arises from the nondetection of SiC in our ISD population, because SiC is not expected to condense in the ISM. Among presolar grains, SiC—condensed in circumstellar atmospheres—should have a 20 to 50% abundance. The 20% estimate is a lower limit, because many meteorites have a SiC/silicate ratio that is higher than 1:5, although this is ascribed to preferential metamorphic destruction of silicates on meteorite parent bodies. On the other hand, model calculations of circumstellar dust (4) also yield relatively high values for the SiC/silicate ratio. The probability of detecting 36 silicates (and no SiC) among a mixture of 80% silicates and 20% SiC is as low as 0.03%, and 5% (i.e., below the 2-sigma confidence level) for an 8% SiC abundance. Destruction and recondensation of grains in the ISM are supported by observations showing that, in the diffuse ISM, the most condensable (atomic mass >23) elements are depleted in the gas phase and hence bound in solids, whereas the average lifetime of interstellar grains against destruction by supernova shocks (about 0.5 billion years) is much shorter than the average residence time of matter in the ISM (2.5 billion years) (4). During this residence time, ISD grains frequently cycle between the hot ISM (low-density regions carved by supernovae), the warm diffuse medium (accessible by spectroscopic observations), and cold molecular clouds, which are star-formation regions. Under the assumption that typically 5% of molecular cloud material is consumed by new stars and planetary systems, interstellar grains cycle on time scales of ~125 million years between the hot medium and the cold medium. Although destruction and devolatilization processes prevail in the hot medium due to high grain velocities and supernova shock waves, recondensation can occur in the cold medium.

Our results indicate that a homogeneous population of grains with roughly average cosmic abundances (concerning Mg:Si:Ca:Fe ratios), possibly including metallic nanophase iron, is the primary constituent of the LIC-ISD, emphasizing the importance of recondensation processes (4, 37). A further implication is that searches for presolar interstellar grains in meteorites led by isotopic anomalies are likely to miss a population that recondensed in the ISM, because destruction and recondensation of solids would erase isotopic anomalies. Hence, a considerable fraction of yet unrecognized—isotopically inconspicuous—ISD could reside in meteorites and cometary material awaiting its discovery [see, for example, the discussion in (41, 42)].

Supplementary Materials

www.sciencemag.org/content/352/6283/312/suppl/DC1

Materials and Methods

Figs. S1 to S6

Table S1

References (4361)

References and Notes

  1. Acknowledgments: N.A. and F.P. acknowledge European Space Agency faculty funding for travels and meetings that were necessary for the completion of this work. Funding through Deutsche Forschungsgemeinschaft (DFG) within the priority program 1385 “The First 10 Million Years of the Solar System - A Planetary Materials Approach” is acknowledged by K.F., J.B. (grant BL 298/20-2), and R.S., M.T., E.G., J.H., and F.P. (grants TR333/14 and SR77/1). J.H. has received funding from the People Programme (Marie Curie Actions) of the European Union’s Seventh Framework Programme FP7 2013 under Reemployment and Eligibility Assessment (REA) grant agreement number 622856. R.S. and F.P. acknowledge funding through Deutsches Zentrum für Luft- und Raumfahrt (DLR), Germany. V.J.S. acknowledges funding through the International Space Science Institute. N.K. acknowledges funding through the DFG. M.T. acknowledges funding by the Klaus Tschira Foundation. We thank J. Leitner, A. Westphal, R. Stroud, L. Nittler, P. Hoppe, H. Ishii, and H.-P. Gail for helpful discussions. We thank three anonymous reviewers for their thorough review of this paper. All CDA data used for this analysis are archived on the Small Bodies Node of the Planetary Data System (PDS-SBN), at http://sbn.psi.edu/archive/cocda.
View Abstract

Navigate This Article